This invention relates to a code multiplexing transmitting apparatus and, more particularly, to a code multiplexing transmitting apparatus for spread-spectrum modulating signals of a plurality of channels by respective ones of codes that differ from one another, combining the spread-spectrum modulated signals of each of the channels and transmitting the resultant spread-spectrum modulated signal.
Wireless access using CDMA (Code Division Multiple Access) has been studied and is being put to use as the next generation of digital mobile communication. CDMA is a method of multiple access using spread-spectrum communication. Specifically, transmission information of a plurality of channels or users is multiplexed by coding and transmitted over a transmission path such as a radio link.
Spread-spectrum communication is a method of modulation that is different from ordinary narrow-band modulation. In spread-spectrum communication, the bandwidth of a signal after modulation is made very large in comparison with that of the narrow band in narrow-band modulation. With spread-spectrum communication, two-stage modulation/demodulation is performed in the transceiver.
FIG. 16 is a structural view illustrating the operating principle of a transmitter in spread-spectrum communication. Shown in FIG. 16 are a modulator 1 such as a (phase-shift keying) PSK modulator, a spreading circuit 2, a power amplifier 3 and an antenna 4. The positions of the modulator 1 and spreading circuit 2 may be interchanged. The spreading circuit 2 includes a spreading code generator 2a for outputting a rectangular spreading code sequence (see FIG. 17) that randomly takes on levels of xc2x11 referred to as a pseudorandom noise (PN) sequence, and a multiplier 2b for multiplying digital transmission data, which has been modulated by the modulator 1, by the spreading code.
As shown in FIG. 17, the speed at which the spreading code changes (namely duration Tc of the rectangular wave) is set so as to change over at a very high rate in comparison with symbol changeover speed (one bit interval T of the PSK-modulated signal) of the narrow-band modulated signal that is modulated by the spreading code. That is, T greater than  greater than Tc holds. The duration of T is referred to as the xe2x80x9cbit durationxe2x80x9d, the duration of Tc is referred to as the xe2x80x9cchip durationxe2x80x9d, and the reciprocals of these are referred to as the xe2x80x9cbit ratexe2x80x9d and xe2x80x9cchip ratexe2x80x9d, respectively. The ratio of T to Tc (i.e. T/Tc) is referred to as the xe2x80x9cspreading ratioxe2x80x9d.
The spectrum distribution of a spread-spectrum modulated signal exhibits the shape of a sinc function, as shown in FIG. 18. The bandwidth of a main lobe ML is equal to twice the chip rate (i.e. ML=2/Tc), and the bandwidth of a side lobe SL is 1/Tc. Since the PSK signal prior to spread-spectrum modulation is an ordinary PSK signal modulated at the bit rate 1/T, the occupied bandwidth is 2/T. Accordingly, if the occupied bandwidth of the spread-spectrum modulated signal is made the bandwidth (=2/Tc) of the main lobe, the bandwidth of the original PSK-modulated signal will be broadened T/Tc times by applying spread-spectrum modulation. The energy is diffused as a result. FIG. 19 is an explanatory view illustrating the manner in which bandwidth is enlarged by spread-spectrum modulation. Shown in FIG. 19 are a narrow bandwidth-modulated signal NM and a spread-spectrum modulated signal SM.
FIG. 20 is a structural view illustrating the operating principle of a receiver in spread-spectrum communication. Shown in FIG. 20 are an antenna 5, a wide-band bandpass filter 6 for passing only signals of necessary frequency bands, a de-spreading circuit 7, a bandpass filter 8 and a detector circuit 9 such as a PSK demodulator. The de-spreading circuit 7 has a construction identical with that of the spreading circuit 2 on the transmitting side and includes a spreading code generator 7a for outputting a rectangular spreading code sequence the same as that on the transmitting side, and a multiplier 7b for multiplying the output signal of the bandpass filter 6 by the spreading code.
The wide-band reception signal sent to the receiver is restored to the original narrow-band modulated signal via the de-spreading circuit 7 similar to the spreading circuit on the transmitting side. This is followed by the generation of a baseband waveform via the detector circuit 9, which is of the ordinary type. The reason why the narrow-band modulated signal is obtained by the de-spreading circuit 7 is as set forth below.
As shown in FIG. 21, let a(t) represent the modulated wave on the transmitting side, c(t) the spreading code sequence (spreading code) and x(t) the transmitted waveform. These are related as follows:
x(t)=a(t)xc2x7c(t)
If attenuation and the effects of noise during transmission are neglected, the transmitted waveform x(t) arrives on the receiving side intact. The spreading code sequence used by the de-spreading circuit 7 has a waveform exactly the same as that of the spreading code used in spread-spectrum modulation on the transmitting side, as mentioned above. Accordingly, the output y(t) of the de-spreading circuit 7 is given by the following equation:
y(t)=x(t)xc2x7c(t)=a(t)xc2x7c2(t)
The output signal y(t) enters the bandpass filter 8. Passing this signal through the bandpass filter is the same as integrating the signal. Thus the output of the bandpass filter is given by the following equation:
xe2x80x83∫y(t)dt =a(t)xc2x7∫c2(t)dt
The integral on the right side of this equation is an autocorrelation value obtained when the shift in time is made zero. The autocorrelation value is unity. Accordingly, the output of the bandpass filter is a(t) and the modulating information signal is obtained.
Code division multiple access (CDMA) is a method of communication using a different spreading code for each channel or user, whereby the information transmitted on the respective channels is multiplexed by the codes. FIG. 22 is a diagram for describing the principle of CDMA on two channels. Shown in FIG. 22 are a transmitter TR in which CH1 is a first channel, CH2 a second channel and CMP a combining unit, and first and second receivers RV1, RV2, respectively.
An important point in CDMA is the xe2x80x9csimilarityxe2x80x9d of the spreading codes used by each of the channels. When almost identical spreading codes are used by each of the channels, the channels interfere with each other severely. A so-called xe2x80x9ccorrelation valuexe2x80x9d is a measure of the degree to which interference between channels occurs. The correlation value is defined by the following equation with respect to two waveforms a(t) and b(t):
R=∫a(t)xc2x7b(t)dt T: period
The integration is carried out over one period T of a(t), b(t). We have R=1 when a(t) and b(t) are exactly identical waveforms and R=xe2x88x921 when the waveforms are of opposite signs. On the average, looking at one period, the value of R obtained is zero when there is no relationship between the value of a(t) at a certain time and the value of b(t) at the same time.
Consider the first receiver RV1 in a situation where CDMA is performed using, as the spreading code, two waveforms c1(t) and c2(t) of such a combination that the correlation value R is zero. The signals from the first and second channels CH1 and CH2 arrive at the first receiver RV1. When the first receiver RV1 de-spreads the received signals using the code c1(t), a bandpass filter 81 outputs a signal represented by the following equation:
∫{a1(t)c1(t)c1(t)+a2(t)c2(t)c1(t)}dt
The ∫{a2(t)c2(t)c1(t)}dt part of this is zero because the correlation value between c2(t) and c1(t) is zero. Further, ∫c1(t)c1(t)dt is unity since this is an autocorrelation value in which the displacement in time is zero. Accordingly, the output of the bandpass filter 81 of the first receiver RV1 is a1(t) and the influence of the signal making use of c2(t) as the spreading code is entirely absent. The same is true for the second receiver RV2. This will hold even if the number of simultaneously connected communication channels is increased. However, it is required that the correlation value be zero for the spreading codes of all combinations.
In mobile wireless communication, wireless base stations emit radio waves (generate spreading code sequences) at the same timing (i.e. synchronously). It will suffice, therefore, to select spreading code sequences in such a manner that the correlation value will be zero between the spreading code sequences. It should be noted that since one wireless mobile station will not emit radio waves at the same timing as other wireless mobile stations, mutual influence cannot be measured merely by the correlation value. Accordingly, the correlation values of c1(t) and c2(t) are not merely compared; it is required that the correlation values be observed for a case where c1(t) and c2(t) are shifted arbitrarily in time.
FIG. 23 is a diagram showing the construction of a CDMA transmitter which code-multiplexes and transmits data on a number of channels. This illustrates the construction of a prior-art base station in wireless mobile communication, by way of example. As shown in FIG. 23, the transmitter includes spread-spectrum modulators 111xcx9c11n of 1st through nth channels, respectively. Each spread-spectrum modulator includes a frame generator 21, a serial/parallel (S/P) converter 22 for converting frame data to parallel data, and a spreading circuit 23. The frame generator 21 has a transmission data generator 21a for generating serial transmission data D1, a pilot signal generator 21b for generating a pilot signal which is peculiar to a base station, and a frame forming unit 21c for forming the serial data D1 (see FIG. 24) into blocks every prescribed number of bits and inserting the pilot signal P at the beginning and end of each block, thereby producing data frames. The frame generators 21 of each of the spread-spectrum modulators 111xcx9c11n insert identical pilot signals P into the transmission data at the same timing. The purpose of the pilot signal P is to allow the receiver to recognize the amount of phase rotation of the spread-spectrum modulated signal due to transmission. In other words, the pilot signals are used to perform de-spreading by allowing the receiver to detect the amount of phase rotation of the spread-spectrum modulated signal in the transmission path from the position of the transmitted pilot and the position of the received pilot, and to restore the phase of the spread-spectrum modulated signal by an amount equivalent to the amount of phase rotation.
The S/P converter alternately distributes the frame data (the pilot signals and transmission data) one bit at a time to convert the frame data to I-component (in-phase component) data DI and Q-component (quadrature-component) data DQ, as shown in FIG. 24.
The spreading circuit 23 includes a pseudorandom noise (pn) sequence generator 23a for generating a pn sequence (long spreading code) which is peculiar to the base station, an orthogonal Gold code generator 23b for generating an orthogonal Gold code (short spreading code) for user identification, an EX-OR gate 23c for obtaining the exclusive-OR between the pn sequence and the orthogonal Gold code and outputting a resulting code C1, and EX-OR gates 23d, 23e for performing spread-spectrum modulation by obtaining the exclusive-ORs between the data DI and DQ, respectively, and the code C1. It should be noted that since xe2x80x9c1xe2x80x9d is level 1 and xe2x80x9c0xe2x80x9d is level xe2x88x921, the exclusive-OR between signals is the same as the product between them.
Also shown in FIG. 23 are a combiner 12i for outputting an I-component code-multiplexed signal xcexa3VI by combining the I-component spread-spectrum modulated signals VI output by the respective spreading circuits 111xcx9c11n; a combiner 12q for outputting a Q-component code-multiplexed signal xcexa3VQ by combining the Q-component spread-spectrum modulated signals VQ output by the respective spreading circuits 111xcx9c11n; FIR-type digital chip shaping filters 14i, 14q for limiting the bandwidths of the code-multiplexed signals xcexa3VI, xcexa3VQ, respectively; DA converters 14i, 14q for converting the digital outputs of the respective filters 13i, 13q to analog signals; a quadrature modulator 15 for applying quadrature phase-shift keying (QPSK) modulation to the code-multiplexed signals xcexa3VI, xcexa3VQ of the I and Q components and outputting the modulated signal; a power amplifier 16 for amplifying the output of the quadrature modulator 15, and an antenna 17.
The quadrature modulator 15 includes a carrier generator 15a for outputting a carrier wave cos xcfx89t having a prescribed frequency, a 90xc2x0 phase shifter 15b for shifting the phase of the carrier wave by 90xc2x0 and outputting xe2x88x92sin xcfx89t, a multiplier 15c for multiplying the output signal of the DA converter 14i by cos xcfx89t, a multiplier 15d for multiplying the output signal of the DA converter 14q by xe2x88x92sin xcfx89t, and a combiner 15e for combining the outputs of the multipliers 15c and 15d. 
FIG. 25 is a diagram showing the construction of the orthogonal Gold code generator 23b. The code generator 23b includes a first M (maximum-length code) sequence generator 23b-1, a second M sequence generator 23b-2, an exclusive-OR gate 23b-3 for obtaining the exclusive-OR between the first and second M sequences, and a xe2x80x9c0xe2x80x9d add-on unit 23b-4 for adding a xe2x80x9c0xe2x80x9d onto the end of the sequence outputted by the exclusive-OR gate 23b-3.
The first M sequence generator 23b-1 has a 6-bit shift register SF1 and an exclusive-OR gate EOR1, generates the M sequence
A={ai, i=0, 1, 2, . . . , Nxe2x88x922}
by performing the operation represented by a primitive polynomial X6+X+1 and adds xe2x80x9c0xe2x80x9d onto the end of the M sequence A, thereby generating a sequence U, of sequence length N=2n, expressed by the following equation:
U=(a0, a1, a2 . . . aNxe2x88x922, 0)=(A,0)
The second M sequence generator 23b-2 has a 6-bit shift register SF2 and an exclusive-OR gate EOR2, generates the M sequence
B={bi, i=0, 1, 2, . . . , Nxe2x88x922}
by performing the operation represented by a primitive polynomial X6+X5+X3+X2+1 and adds xe2x80x9c0xe2x80x9d onto the end of the M sequence B, thereby generating a sequence Vj, of sequence length N=2n, expressed by the following equation:
Vj=[Tj(b0, b1, b2 . . . bNxe2x88x922), 0]=(TjB,0)
where TjB is the result of shifting the sequence B by j. The orthogonal Gold code is produced from the sequences U, Vj and is composed of a set of N sequences.
The first M sequence generator 23b-1 generates the sequence U (the initial value of the shift register SF1 being made 000001). The second M sequence generator 23b-2, on the other hand, generates the sequence B with xe2x80x98000000xe2x80x99 being the initial value of the shift register SF2, and generates the sequence Vj by shifting the sequence B (Nxe2x88x921) times. Next, the exclusive-OR gate 23b-3 obtains the exclusive-OR between the sequences U and Vj and outputs (Nxe2x88x921) items of data. After the (Nxe2x88x921) items of data are output, the xe2x80x9c0xe2x80x9d add-on unit 23b-4 outputs xe2x80x9c0xe2x80x9d as the N-th item of data, thereby generating a first orthogonal code sequence G1.
Next, the first M sequence generator 23b-1 generates the sequence U (the initial value of the shift register SF1 being made 000001). The second M sequence generator 23b-2, on the other hand, generates the sequence B with xe2x80x98000000xe2x80x99 being the initial value of the shift register SF2, and generates the sequence Vj by shifting the sequence B (Nxe2x88x922) times. Next, the exclusive-OR gate 23b-3 obtains the exclusive-OR between the sequences U and Vj and outputs (Nxe2x88x921) items of data. After the (Nxe2x88x921) items of data are output, the xe2x80x9c0xe2x80x9d add-on unit 23b-4 outputs xe2x80x9c0xe2x80x9d as the N-th item of data, thereby generating a second orthogonal code sequence G2.
Thereafter, and in similar fashion, (Nxe2x88x922) sequences G3xcx9cGN are generated. As a result, a set of a total of N sequences G1xcx9cGN is obtained. A feature of these codes is orthogonality between the code sequences. FIG. 26 shows an example of 64 orthogonal Gold code sequences, each having a code length of 64 bits, generated in the manner described above. The last value of each sequence is xe2x80x9c0xe2x80x9d.
A multiplexed signal of pilots in a case where code multiplexing has been performed using the above-mentioned orthogonal Gold codes with pilots in phase is expressed as follows, where the data dealt with is (xe2x88x921, +1):                               MultiCode          pilot                =                ⁢                              ∑            i            user                    ⁢                      xe2x80x83                    ⁢                      (                                          ogold                i                            xc3x97              Pilot              xc3x97              PN                        )                                                  =                ⁢                  Pilot          xc3x97          PN          xc3x97                                    ∑              i              user                        ⁢                          (                              ogold                i                            )                                                              =                ⁢                  C          xc3x97                                    ∑              i              user                        ⁢                          (                              ogold                i                            )                                          
Consider the right side of this equation. The amplitude of the multiplexed signal takes on the maximum value at the portion where xe2x80x9c0xe2x80x9d is given as the Nth item of data when the orthogonal Gold codes are generated (xe2x80x9c0xe2x80x9d corresponds to the xe2x88x921 level), as shown in FIG. 27. The reason for this is that since the amplitude (the outputs of the combiners 12i, 12q in FIG. 24) of a multiplexed signal in CDMA is the sum of the voltages of all multiplexed channels, the maximum value is obtained when the orthogonal Gold codes are all xe2x80x9c0xe2x80x9ds or all xe2x80x9c1xe2x80x9ds.
Thus, in pilot-insertion type CDMA, pilot signals are added on frame by frame and the pilot signals are spread-spectrum modulated by orthogonal codes (orthogonal Gold codes) for user identification and a pn sequence. Let n represent the number of channels. After code multiplexing n-number of spread-spectrum modulated signals that have been generated, a CDMA base station applies QPSK modulation and then transmits the modulated signal. When the n channels of spread-spectrum modulated signals are code-multiplexed in such a CDMA base station, the pilot signals are in common for each of the channels and the output timings of the pilot signals of each of the channels are the same. Consequently, the power of the signal obtained by n-code multiplexing the spread-spectrum modulated signals develops peak values at the points where the pilot signals reside, as shown in FIG. 28. This is a problem in that these peaks of the multiplexed signal act as interference waves with respect to other stations.
Another factor is that the input/output characteristic of a power amplifier is linear up to a certain input level but becomes non-linear when this level is exceeded. FIG. 29 shows an example of an AM-AM characteristic (input power vs. gain characteristic) of a power amplifier, and FIG. 30 shows an example of an AM-PM characteristic (input power vs. phase characteristic) of a power amplifier. It will be understood from these characteristic curves that the gain characteristic and phase characteristic of a power amplifier are flat and so is the input/output characteristic as long as the input power is small. There is also no phase rotation under these conditions. However, when the input power exceeds a certain level, gain starts to decline, a phase lag develops and each characteristic becomes non-linear. It is required to use a power amplifier with a high power efficiency and it is necessary to raise the mean power level of the input signal. When the mean power level of the input signal is raised, however, the peak value of the code-multiplexed signal exceeds the linear region and saturates and the peak values at the locations of the pilot signals are clipped, as shown in FIG. 31. As a result, when this code-multiplexed signal is de-spread on the receiving side, the pilot signal power becomes small in comparison with the power of the other data, pilot detection error increases and the amount of phase rotation can no longer be recognized. The result is that data can no longer be demodulated correctly. If the mean power level of the input signal is used upon being reduced, a problem which arises is a decline in the power efficiency of the power amplifier.
Accordingly, an object of the present invention is to reduce the peak values of identical signal portions, e.g. the pilot signal portions, of identical timings in the code-multiplexed signal.
Another object of the present invention is to make it possible to reduce the power of radio waves which interfere with other stations, thereby increasing system capacity.
A further object of the present invention is to make efficient use of the power amplifier.
In accordance with the present invention, the foregoing objects are attained by providing a code multiplexing transmitting apparatus for spread-spectrum modulating transmission data of a plurality of channels by spreading codes that differ from one another, combining the spread-spectrum modulated signals of each of the channels and transmitting the resultant spread-spectrum modulated signal, comprising a phase shifter for shifting, by a predetermined angle channel by channel, phase of a signal-point position vector of the spread-spectrum modulated signal of each channel.
Other features and advantages of the present invention will be apparent from the following description taken in conjunction with the accompanying drawings.